Contents lists available at ScienceDirect

Land Use Policy

journal homepage: www.elsevier.com/locate/landusepol

Bits and pieces: Forest fragmentation by linear intrusions in IndiaRajat Nayaka,*, Krithi K. Karanthb,c,d, Trishna Duttae, Ruth Defriesf, K. Ullas Karanthb,c,g,Srinivas Vaidyanathanaa Foundation for Ecological Research, Advocacy and Learning, 170/3, Morattandi, Tamil Nadu, 605101, IndiabWildlife Conservation Society, Bronx, New York, 10460, USAc Centre for Wildlife Studies, 2nd Phase Kodigehalli, Bengaluru, 560097, IndiadNicholas School of the Environment, Duke University, Durham, North Carolina, USAeWildlife Sciences, Faculty of Forest Sciences and Forest Ecology, University of Goettingen, Goettingen, 37077, GermanyfDepartment of Ecology, Evolution, and Environmental Biology, Columbia University, New York, NY, 10027, USAgWildlife Conservation Society, India Program, Bengaluru, 560097, India

A R T I C L E I N F O

Keywords:InfrastructurePatch indicesPatch sizeInter-patch distanceWestern GhatsCentral India

A B S T R A C T

Linear infrastructure development is an important driver of forest fragmentation leading to habitat and biodi-versity loss as well as disruption of critical ecosystem processes. The tropical forests of India are increasinglyimpacted by infrastructure development. Little quantitative information is available on the extent of fragmen-tation due to linear infrastructure on these habitats. Here, we quantified fragmentation due to linear infra-structure by studying forest structural connectivity. We compared the existing forest patch characteristics with ascenario that excluded all linear infrastructure. We classified forest patches into three different fragmentationcategories that combined information on patch size, inter patch distance and percentage perforations. Resultsshow that power-transmission lines and roads were the most common infrastructure features within forests. Wefound a 6% increase in the number of forest patches due to the construction of linear infrastructure. Forestpatches> 10,000 km2 in size were severely affected and there was a 71.5 % reduction in the number of suchpatches. We found that 86 % of the existing forest patches are in the small (median patch size< 1 km2) andisolated (a median distance of 155m) category. The density of linear infrastructure inside protected areas wassimilar to density in non-protected forested areas. Our results highlight the need to minimize the effects offragmentation in the future by considering re-routing or bundling of infrastructure. When infrastructure is un-avoidable, there is a need to mitigate their potential impacts. The results of this study have been made publiclyaccessible (https://indiaunderconstruction.com) to provide information on 'where' to avoid future linear infra-structure development and to make informed decisions which can lead to optimally designed local managementplans.

1. Introduction

Tropical forests are one of the most diverse ecosystems in the world.They are also one of the most threatened ecosystems undergoing rapidland use changes and fragmentation (Achard et al., 2002; Gibson et al.,2011; Hill et al., 2011; Laurance and Bierregaard, 1997; Mayaux et al.,2005; Miles et al., 2006; Ramachandran et al., 2018). One of the im-portant factors contributing to fragmentation in tropical forests is in-frastructure development (Geist and Lambin, 2002; Geneletti, 2004;

Goosem, 1997, 2007; Laurance, 2015; Laurance et al., 2014; Reed et al.,1996). Infrastructure, especially linear structures such as roads, railwaylines, power-transmission lines, canals, and pipelines create linear gapswhich split a contiguous forested area into smaller units known aspatches (Geneletti, 2004; Hawbaker et al., 2006; Mancebo Quintanaet al., 2010; Miller et al., 1996)

Linear fragmentation leads to a reduction in habitat area and in-creased habitat isolation, which in turn affects biodiversity and wildlifemovement across forests (Goosem, 2007; Karlson and Mörtberg, 2015;

https://doi.org/10.1016/j.landusepol.2020.104619Received 4 September 2018; Received in revised form 17 March 2020; Accepted 18 March 2020

Abbreviations: AP, amount of perforation; ENVIS, environmental information system; FC, fragmentation category; India-WRIS, water resources information systemof India; IPD, inter-patch distance; LULC, land use/land cover; MoEFCC, Ministry of Environment Forest and Climate Change; NRSA, National Remote SensingAgency; PA, protected area; PS, patch size

⁎ Corresponding author at: Foundation for Ecological Research, Advocacy and Learning, 170/3, Morattandi, Tamil Nadu, 605101, India.E-mail addresses: rajat@feralindia.org (R. Nayak), krithi.karanth@gmail.com (K.K. Karanth), trishnad@gmail.com (T. Dutta), rd2402@columbia.edu (R. Defries),

ukaranth@gmail.com (K.U. Karanth), srinivasv@feralindia.org (S. Vaidyanathan).

Land Use Policy xxx (xxxx) xxxx

0264-8377/ © 2020 Elsevier Ltd. All rights reserved.

Please cite this article as: Rajat Nayak, et al., Land Use Policy, https://doi.org/10.1016/j.landusepol.2020.104619

http://www.sciencedirect.com/science/journal/02648377https://www.elsevier.com/locate/landusepolhttps://doi.org/10.1016/j.landusepol.2020.104619https://indiaunderconstruction.comhttps://doi.org/10.1016/j.landusepol.2020.104619mailto:rajat@feralindia.orgmailto:krithi.karanth@gmail.commailto:trishnad@gmail.commailto:rd2402@columbia.edumailto:ukaranth@gmail.commailto:srinivasv@feralindia.orghttps://doi.org/10.1016/j.landusepol.2020.104619

Lovejoy et al., 1986). The problem is acute in tropical forests wherecertain animals tend to avoid forest clearings as narrow as 30m (reviewin Laurance et al., 2009). Infrastructure act as barriers to faunalmovement and affect habitat use and migration paths (Bhattacharyaet al., 2002; Develey and Stouffer, 2001; Forman et al., 1997; Kocioleket al., 2011; Shepard et al., 2008; Strand, 2004). This barrier effect bylinear structures may also reduce gene flow (Riley et al., 2006) andaffect associated population sizes and densities (Benítez-López et al.,2010; Lesbarrères and Fahrig, 2012). Mortality of wildlife due to roadkills and electrocution is well documented along roads, railways, andpower-transmission lines (Coffin, 2007; Jaarsma, 2006; W. F. Lauranceet al., 2009; Uddin, 2017). The patches resulting from linear gaps maybe too small with limited resources and detrimental for the survival ofsome species, resulting in reduced diversity (Coffin, 2007; Fahrig, 2002;Girardet et al., 2013; Goosem, 2007; Mancebo Quintana et al., 2010;Opdam et al., 2001). Apart from the barrier effect and impacts onbiodiversity, linear infrastructure can change soil properties, hydrologiccycles, and other ecosystem processes and functioning, and facilitatedispersal of invasive and pathogens into natural habitats (Forman andAlexander, 1998; Laurance et al., 2009).

India is one of the 17 ‘mega-diverse countries’ in the world knownfor high endemism (Mittermeier et al., 2004). Most of the endemicspecies are restricted to subtropical and tropical forested tracts whichconstitute nearly 23 % of India’s geographical area. In the recent pastIndia’s economy has witnessed a fast growth. Indian infrastructurenetwork, especially roads, railway, and power-lines, have been greatlyexpanded and upgraded. For example, the increase in the length ofhighways (national and state highways) between 1980 and 2000 was50 %, while this length was increased by nearly 40 % between 2001 and2015 (https://data.gov.in/). In the process, many infrastructure pro-jects were undertaken in pristine forested habitats and there have beennearly 7000 linear project proposals submitted to the Government ofIndia for forest clearance between July 2014 and September 2017(MoEFCC 2017).

Information on the distribution of forest patches and their con-nectivity is essential to account for biodiversity conservation in infra-structure development at a national level (Seiler and Eriksson, 1997).Recently, there has been growing interest in integrating conservationconcerns in infrastructure developments in India (Dutta et al., 2018;WII, 2016). Although there is enough evidence on the impacts of linearinfrastructure on forests, there is an urgent need to illustrate the extentof forest fragmentation due to linear infrastructure in India.

The main objective of this study was to understand the impact ofinfrastructure developments on forest structural connectivity in India.In this study, we quantified structural connectivity, which is defined asthe spatial arrangement of forested habitats in a landscape, by ana-lyzing forest patch characteristics. We quantified the impact of infra-structure development on structural connectivity by comparing existing

forest patch characteristics that take into account the entire linear in-frastructure, with a scenario that excluded linear intrusions withinforested areas. We used patch indices – patch size, amount of perfora-tion and inter-patch distance, to quantify fragmentation using GIS.Reduction in patch size and increase in the number of patches is di-rectly linked to the splitting of a large forest habitat by linear structures(Goosem, 2007; Lovejoy et al., 1986; Mancebo Quintana et al., 2010),while an increase in inter-patch distance and perforation is a combi-nation of the direct effect of linear infrastructure and associatedchanges in land use and land cover and increased human activity(Laurance et al., 2009). We performed cluster analysis using patch in-dices to identify large, intact patches that need to be preserved in futuredevelopment action plans. We summarise our results at the nationalscale and for the existing protected area (PA) network. The results arealso presented for two important conservation landscapes; the WesternGhats and Central India, which are rich in biodiversity and critical forsurvival of several threatened large mammals including tiger (Pantheratigris) and Asian elephant (Elephas maximus) (Baskaran et al., 1995;Baskaran, 2013; Sanderson et al., 2006). Both of these landscapes arealso under severe development pressure from linear infrastructureprojects. We present the most comprehensive map of structural con-nectivity and fragmentation for India which takes into account variousinfrastructure features.

2. Methods

2.1. Datasets

We used open access spatial datasets available in the public domain,updated up to 2015, for our analysis. We used datasets hosted andverified by government agencies, such as National Remote SensingAgency (NRSA), Water Resources Information System of India (India-WRIS), Indian Rail Information System, National Highway Authority ofIndia, and Ministry of Environment Forest and Climate Change(MoEFCC). We also made use of datasets from OpenStreet Maps (OSM)which were verified and updated up to 2015 using Google Maps.

2.1.1. Forest cover and infrastructure layersThe forest cover layer was derived from land use/land cover (LULC)

map provided by the NRSA for the year 2014−15. We considered fourmajor linear infrastructure features in this study; roads, canals, rail-ways, and high tension power-lines. Apart from linear infrastructure,we used a reservoir and mines/quarry layer, which are known sourcesof forest cover loss and fragmentation. The source and details of theselayers are provided in Supplementary Table 1.

2.1.2. Administrative and landscape boundariesWe obtained boundary files of the States and Union Territories,

Table 1A summary of patch size (PS), inter-patch distance (IPD) and amount of perforation (AP) across the clusters and fragmentation categories (FC). FC have beengenerated for forest patches in the mainland India and islands were excluded from cluster analysis due to their large IPD values.

Cluster Number of Patches PS (km2) IPD (m) AP (%) Fragmentation Category (FC)

1 1 60,500.0 55.0 3.64 FC 12 5086 Median: 3.41 Median: 55.0 Median: 6.41 FC 1

Range: Range: Range:0.02–16,849.0 55.0–7,880.0 0.1–43.7

3 302,482 Median: 0.05 Median: 155.0 Median: 0.00 FC 2Range: Range: Range:0.009–412.0 55.0–2,709.0 0.0–4.86

4 2339 Median: 0.04 Median: 5,903.0 Median: 0.00 FC3Range: Range: Range:0.009 – 11.0 4,338.0–56,656.0 0.0–4.8

5 42,448 Median: 0.03 Median: 1107.0 Median:0.00 FC3Range: Range: Range:0.009 – 27.0 676.0–4,331.0 0.0–2.14

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

2

https://data.gov.in/

which are the main administrative units under the Indian federalsystem, from Survey of India (Fig. 1). There is no accurate PA layeravailable in the public domain. The widely used world database onprotected areas is incomplete and inaccurate for India. Hence we col-lated PA boundaries across different government and published sources:PA layer for the Western Ghats was obtained from the dataset generatedby Das et al., 2006. Boundaries of other PAs were digitized from thegazette notifications available with ENVIS PA database (ENVIS Centreon Wildlife & Protected Areas: http://www.wiienvis.nic.in/Database/Protected_Area_854.aspx) and from Eco-Sensitive Zones notificationsavailable with the MoEFCC (http://envfor.nic.in/content/esz-notifications). This PA layer was further updated with input from re-searchers working in different landscapes in India.

We used the Western Ghats boundary as defined by Das et al., 2006.It demarcates an area of 120,000 km2of hill range running through alength of 1600 km from north to south along the west coast of India,covering six states, as the Western Ghats (Fig. 1). We delineated CentralIndia boundary using agro-ecological regions and district boundaries.We used four agro-ecological regions, a) Deccan Plateau hot semi-aridecoregion, b) Central Highlands hot sub-humid ecoregion, c) Eastern

Plateau (Chattisgarh) hot sub-humid ecoregion, and d) Eastern Plateau(Chhota Nagpur) and Eastern Ghats hot sub-humid ecoregion (Sehgalet al., 1992). All the districts falling within these zones across ninestates, Madhya Pradesh, Maharashtra, Chattisgarh, Jharkhand, Tel-angana, Rajasthan, Uttar Pradesh, Bihar, and Orissa, covering an area of729,000 km2, were selected to delineate the boundary.

2.2. Preprocessing datasets

The LULC with 19 classes was re-sampled to 55m using the nearestneighbour method and converted to a forest only layer by extracting theclasses - Evergreen Forest, Deciduous Forest, Scrubland, and Littoralinto one single “forest” class. Before extracting forest only class, weremoved all stray pixels which resulted in a small isolated island likeclasses within a larger class, by merging them into the surroundinglarger class using r.neighbors module in GRASS 7.2, with five neigh-borhood cells and a mode function. Some seasonal rivers inside forestswere classified as “water-body” in the original LULC layer. As thesewere seasonal water bodies with a narrow width and known to allowmovement of plants and animals across in the drier periods, we

Fig. 1. Map showing administrative boundaries of India, protected areas and the two landscapes, the Western Ghats and Central India.

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

3

http://www.wiienvis.nic.in/Database/Protected_Area_854.aspxhttp://www.wiienvis.nic.in/Database/Protected_Area_854.aspxhttp://envfor.nic.in/content/esz-notificationshttp://envfor.nic.in/content/esz-notifications

reassigned such pixels to the “forest” class. A river layer, obtained fromthe India-WRIS web-service, and surface water occurrence/seasonalitylayer obtained from global surface water explorer (Pekel et al., 2016)was used to facilitate this process.

We checked the topology of all vector layers (roads, railways,power-lines, mines, and reservoirs and dams) and errors such as over-laps, gaps and duplicate geometries were cleaned and the topology wasfixed. Further, missing attribute values were updated. The road layerwas restricted to major roads - national highways, state highways, anddistrict roads, as other roads such as rural and urban roads were notcaptured effectively by the available road database. This layer wasfurther updated by digitising missing roads using Google Earth scenes,available for the years 2014 and 2015. The railway line layer was up-dated and refined using the Indian Rail Atlas (https://indiarailinfo.com/atlas). The raster layer of canals and reservoirs was converted intovector layers. The quarry/mines layer was further updated by digitisingopen cast mines from Google Earth scenes for the Western Ghats andCentral India available for 2014−15.

We used the r.clump module in GRASS 7.2 to identify physicallydiscrete forest patches and converted the resulting raster layer into avector polygon layer. The original LULC classification does not ade-quately represent linear infrastructure within forested areas; ideally,these features should have been classified as built-up (roads andrailway line) or as water bodies (canals). To address this under-re-presentation, we used infrastructure vector layers obtained from dif-ferent sources and added a buffer of 27.5m on either side of all linearfeatures to match the spatial resolution of the LULC layer. We thenconducted a vector difference operation and deducted all infrastructurefeatures, linear as well as reservoir and mines, from the forest patchlayer. We further removed patches which were less than 3 pixels in size(< 9000 m2) from the resulting layer. This was our final layer whichshowed all distinct forest patches in India accounting for all infra-structure (Fig. 1). At the end of this exercise, we had two forest coverlayers; one representing the forest patches without linear intrusions(simulated forest area) and one with all linear infrastructure included(henceforth existing forest area).

2.3. Analysis

We derived three different patch indices for the existing forest coverto quantify the structural connectivity:

Patch Size (PS): This is a simple measure of the size of the patchobtained by measuring the geometrical area of each individual patch.The area was expressed in m2. Reduction in patch size is directly linkedto the splitting of a larger forest habitat into smaller parts by linearstructures (Goosem, 2007; Lovejoy et al., 1986; Mancebo Quintanaet al., 2010). There is also an association between patch size and its useby animals (Uezu et al., 2005; Webb, 2013). Hence we chose this indexfor quantifying structural connectivity.

Amount of Perforation (AP): This is a measure of gaps or holes in aforest patch, which can be created by reservoirs, mines, settlements,agriculture/non-forest plantations, or changes in LULC associated withinfrastructure developments, expressed as a percentage of the totalpatch size.

Inter-Patch Distance (IPD): This is an index of isolation of a forestpatch in space from the nearest forest patch. IPD was defined as thenearest neighbor distance (m) for each forest patch. Patch shrinkage bylinear intrusions could lead to an increase in distance between patches.If such distances are very large it could result in the complete isolationof a patch from the other and thus impact movement patterns and co-lonization of animals and plants across patches and landscapes(Schowalter, 2016; Webb, 2013). Hence we used this index in combi-nation with PS and AP to quantify the impact of linear infrastructure onstructural connectivity.

Patch size and the number of patches were compared between ex-isting forest area and simulated forest area layers to evaluate the impactof infrastructure development on forest structural connectivity.

We performed agglomerative hierarchical cluster analysis on patchindices to identify naturally distinct clusters in our dataset. We usedWard’s method of agglomeration, which produces clusters of moreequal size by keeping distances within the clusters as small as possible(Ward, 1963). We identified five unique groups using a dendrogram onthe cluster solutions that we had obtained (Supplementary Fig. 1).These five groups were further classified into three fragmentation ca-tegories (FC) based on the patch characteristics (Table 1). FC rangedfrom small and distant patches to large and intact patches. This formedthe basis for quantifying fragmentation and recommending planningpriorities for infrastructure developments in the study region. We didnot include island landscapes such as Andaman and Nicobar, andLakshadweep Islands in the cluster analysis due to the large IPD valuesof forested patches. Cluster analysis was performed using R statisticalsoftware (R Core Team, 2018) and the hclust function available with Rlibrary fastcluster (Müllner, 2013).

We used infrastructure information and patch indices to char-acterize fragmentation in existing PAs and in Central India and theWestern Ghats landscapes.

3. Results

We found an increase in the number of forest patches and a re-duction in the number of large patches (> 10,000 km2) due to linearinfrastructure in India. High tension power-transmission lines andmajor roads were the most common linear intrusions within forests, and70 % of the assessed PAs had some amount of linear infrastructurepassing through them. Forest patches in Central India were more iso-lated than patches in the Western Ghats landscape.

3.1. Infrastructure and patch indices at the national level

We found an increase in the number of forest patches as a result ofinfrastructure construction in India. The total existing forest area of783,300 km2, 23.83 % of India’s landmass, is distributed across352,674 forest patches of varying sizes. High tension power-transmis-sion lines and major roads were the most common linear intrusionswithin forests with lengths of 59,500 km (density=0.08 km/km2) and46,700 km (density=0.06 km/km2), respectively. The length of railwaylines and canals passing through forests were 7400 km (den-sity=0.01 km/km2) and 6100 km (density=0.008 km/km2) respec-tively. A comparison of the two forest layers suggested a 6% increase inthe number of forest patches (patches without infrastructure=331,240, patches with infrastructure= 352,674).We found that approximately 18,250 large forest habitats under the

simulated scenario were split into two or more smaller patches by linearintrusions across the country. Larger forest habitats faced greaterfragmentation and were split into multiple smaller patches by linearinfrastructure intrusions (Fig. 2a). The highest fragmentation due tolinear intrusions was observed in Central India, where an intact foresthabitat of size 162,000 km2 was split into 5200 smaller patches with amean patch size of 30 km2and the largest patch being 16,850 km2insize.

Reduction in size was observed in forest habitats that were impactedby linear intrusions. Eighty-one percent of forest habitats in the simu-lated scenario were reduced by half or more of their original size due tolinear infrastructure (Fig. 2b). A comparison of median patch areasuggested that forest patches of size> 10 km2 were more vulnerable tosize reduction with greater than 50 percent reduction in the patch sizeobserved under simulated scenario (Fig. 2c). The largest patch size was60,500 km2 observed in the North Eastern part of the country, spread

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

4

https://indiarailinfo.com/atlashttps://indiarailinfo.com/atlas

across the states of Arunachal Pradesh, Nagaland, and Manipur. Thissame patch was part of a larger patch of size 172,820 km2 covering allseven north-eastern states in the simulated scenario, which was splitinto 1656 smaller patches with a mean patch size of 100 km2 (Fig. 3 a&b). Furthermore, there was a 71.5 % reduction in the number of largepatches (> 10,000 km2) due to linear infrastructure.

Most of the forest patches (> 94 %), were less than 1 km2 in size.However, they only accounted for approximately 4% of the totalforested area in the country under both existing and simulated forestcover conditions. The distribution of existing forested areas acrossdifferent patch size classes suggested that around 68 % of the forestarea is comprised of forest patches of size between 250 and 10,000 km2

(Fig. 3c). This forest area distribution across patch sizes changes dras-tically in absence of intrusions; forest patches of size> 10,000 km2

alone covered nearly 67 % of the forested area in the simulated scenario(Fig. 3d).

The results for perforation analysis suggested that more than 98 %of forest patches had less than 5% perforation (Fig. 4a). There were only91 forest patches which had more than 30 % perforation. The maximumperforation observed was 43.7 %, calculated for a patch of size 280 km2

in Himachal Pradesh. A comparison of patch size to the amount ofperforation suggested that the perforation was higher for patch sizeclass 1,000–10,000 km2 (median 8.03 %, Fig. 4b). The largest patchwithout any perforations was 77 km2 observed in the state of Gujarat.

Around 92 % of forest patches had an IPD less than 1 km with 33 %having a distance less than 100m (Fig. 4 c&d).

3.2. Fragmentation category (FC)

We found 5087 large, intact forest patches in FC 1 across India,which together formed 77 % of the total forested area in India. Eighty

six percent of the total numbers of forest patches were in FC 2 (Table 1,Fig. 5).

3.3. Status of existing Protected Area (PA) Network

There are 769 PAs in India (ENVIS Centre on Wildlife & ProtectedAreas: http://www.wiienvis.nic.in/Database/Protected_Area_854.aspx). We were able to characterize fragmentation for 450 of the PAswhich predominantly had forest cover. The rest of the PAs were pri-marily grasslands, water-bodies, snow-clad mountains, or marine pro-tected areas and therefore were not assessed. Seventy percent of theassessed PAs had some amount of linear infrastructure passing throughthem. High tension power-transmission lines and major roads were themost common linear intrusions inside PAs with total lengths of12,000 km (density=0.06 km/km2) and 10,000 km (density=0.05 km/km2) respectively. Canals and railways had lengths of 2800 km (den-sity=0.02 km/km2) and 1270 km (density=0.007 km/km2), respec-tively, inside PAs. The densities of high tension power-transmissionlines, major roads, railway lines, and canals were 0.08, 0.06, 0.01, and0.008 respectively across all forested areas in India.

We found around 13,088 forest patches within the PA network. Theaverage patch size within PAs was 30.80 km2. We found that the ma-jority of the forest patches within PAs, 70 %, were part of FC 1, whichincludes large and relatively intact forest patches.

3.4. A comparison between the Western Ghats and Central India

The existing forest area in Central India and the Western Ghatslandscapes were 35 % (238,000 km2) and 68 % (79,900 km2), respec-tively. Nearly 25 % of the forested area in the Western Ghats was insidethe PA network, while only 11.34 % was within the PA network in

Fig. 2. (a) The log plot showing the relationship betweennumber of forest fragments and the initial stage patch size(simulated scenario), which suggests an increase in number offragments with an increase in patch size, (b) A greater re-duction in size was observed in intact forest patches. 81 % ofthese patches were reduced by half or more of their size. (c)Box and whisker plot suggests that intact forest patches ofsize> 1 km2 were more vulnerable to fragmentation with>50 % reduction in original size.

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

5

http://www.wiienvis.nic.in/Database/Protected_Area_854.aspxhttp://www.wiienvis.nic.in/Database/Protected_Area_854.aspx

Central India. The density of roads was 0.12 km/km2 and 0.22 km/km2

inside Central India and the Western Ghats respectively. We found69,458 distinct patches in Central India and 20,284 patches in theWestern Ghats. The mean patch size in the Western Ghats was 3.94 km2,whereas the mean patch size in Central India was 3.42 km2.Comparison of the distribution of forest areas across patch sizes sug-gests that nearly 300 patches of size greater than 1000 km2constituted∼42 % of the forest area in Central India landscape, whereas in theWestern Ghats landscape 14 large patches (PS > 1000 km2)formed∼29 % of the forest area (Fig. 6 a&b). The largest patch size was16,850 km2 and 2940 km2 in Central India and the Western Ghats re-spectively. Comparison of the median AP between Central India and theWestern Ghats landscapes suggested an increase in AP with an increasein PS. Median AP for larger patches of size> 250 km2 was higher in

Central India than in the Western Ghats landscape (Fig. 6 c&d). Forestsin Central India were more isolated than the Western Ghats withmaximum and mean IPD in Central India being 28.4 km and 340mrespectively, while the maximum and mean IPD were 4.6 km and 170mrespectively for the Western Ghats. Similarly, nearly 62 % of patches inthe Western Ghats landscape had IPD of 100m or less, while only 48 %of patches in Central India had IPD of 100m or less. There were 1240and 598 large, intact forest patches (FC 1) in Central India and theWestern Ghats respectively.

4. Discussion

Infrastructure development is the main cause for the fragmentationand loss of connectivity in tropical forests (Geneletti, 2004). In this

Fig. 3. Spatial distribution of forest patches and their size: (a) Depicts the patch size distribution as influenced by infrastructure; (b) Depicts patch size distribution inabsence of infrastructure intrusion. A comparison of both figures suggests higher numbers of large intact patches in the absence of infrastructure intrusions acrossIndia. Distribution of patch size and % of total forested area covered under different patch size: (c) patch size distribution as influenced by infrastructure. Around 58% of total forested area is composed of patches of size 250-10,000 km2; (d) Patch size distribution in absence of infrastructure intrusions. In absence of infrastructureintrusions, around 67 % of forested area is covered by only 7 large patches of size> 10,000 km2.

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

6

study, we quantified the effect of linear infrastructure on forest struc-tural connectivity by using patch characteristics in India. We found anincrease in the number of forest patches as a result of linear infra-structure. Linear infrastructure also resulted in a reduction in the size ofthe forested habitats.

Power-lines and roads were the most common linear intrusionsobserved within forested habitats. The density of roads within forestedhabitats (0.6 km/km2) in India is comparable to average road densitiesreported for some of the developed countries such as USA (0.75 km/km2, Forman, 2003) and higher than that reported for developedcountries such as New Zealand (0.35 km/km2) and Navarra in Spain(0.45 km/km2, Serrano et al., 2002), and for developing countries suchas China (0.43 km/km2) and Brazil (0.20 km/km2) (https://knoema.com/atlas/ranks/Road-density), which have an economic developmentsimilar to India. Our estimates are conservative as we have consideredonly major roads (national and state highways, and district roads) inthis analysis. As there has been an increase of 69 % in the length of ruralroads between 2000 and 2015 (https://data.gov.in/), the actual densityand fragmentation by roads would be much higher than reported herewhen rural and other roads are included in the analysis. Although theserural roads might not be a complete barrier to wildlife movement, theymay result in increased mortality, low patch quality, and a higher edgeeffect for species sensitive to habitat change. With proposed centrally-

sponsored schemes like Bharatmala Pariyojana– a road and highwaydevelopment project (National portal of India 2018, https://www.india.gov.in/spotlight/bharatmala-pariyojana-stepping-stone-towards-new-india), and industrial corridor project along the existing highways(Makeinindia 2018, http://www.makeinindia.com/live-projects-industrial-corridor), the density of roads through forests is likely toincrease in future. This can lead to an increase in fragmentation asfragmentation is closely associated with an increase in road density(Hawbaker et al., 2006). Also, there is potential for a future increase inperforation and inter-patch distance due to the diversion of forest landto non-forest activities.

Up-gradation of road lanes will potentially increase road width andthis can strongly hinder animal movements due to heavier traffic vo-lume and speeds. Laurance and Gomez (2005) reported the inability oftranslocated male Amazonian understory birds to cross clearings of awidth of 250m. Similarly, tigers, snakes, turtles, bumblebees and sev-eral other vertebrates and invertebrates were found to demonstrate astrong avoidance of roads and railway lines (Bhattacharya et al., 2002;Kerley et al., 2002; W. F. Laurance et al., 2017; Shepard et al., 2008).Although the amount of forest land diverted to linear infrastructure issmall, the effect of infrastructure spreads far beyond the exact locationof these structures. A meta-analysis on the effects of infrastructureproximity on mammal and bird populations suggest a decline in

Fig. 4. (a) Map showing the amount of perforation across forest patches. (b) The median for amount of perforation is higher in the patch size class 1000-10,000 km2.(c & d). Majority of patches had less than 1 km inter-patch distance, with 33 % of patches having a distance of less than 100m.

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

7

https://knoema.com/atlas/ranks/Road-densityhttps://knoema.com/atlas/ranks/Road-densityhttps://data.gov.in/https://www.india.gov.in/spotlight/bharatmala-pariyojana-stepping-stone-towards-new-indiahttps://www.india.gov.in/spotlight/bharatmala-pariyojana-stepping-stone-towards-new-indiahttps://www.india.gov.in/spotlight/bharatmala-pariyojana-stepping-stone-towards-new-indiahttp://www.makeinindia.com/live-projects-industrial-corridorhttp://www.makeinindia.com/live-projects-industrial-corridor

abundances of birds and mammals within 2.6 km and 17 km from in-frastructure, respectively (Benítez-López et al., 2010). Although recentmeta-analysis studies suggest a positive effect of human paths onoverall plant species richness and diversity (Root‐Bernstein andSvenning, 2018), the richness and abundance of native plant specieshave been found to decline over distances up to 1 km from roads(Gelbard and Harrison, 2003). Hence there is a need to integrate con-servation concerns in infrastructure development projects and our datacould act as a baseline data for planning and rerouting infrastructureprojects to minimize fragmentation.

We found 352,674 forest patches in India, similar to that reportedby Reddy et al. (2013) (370,386). Although this is a large number,nearly 60 % of India’s existing forest cover is represented by less than1% of these patches. Fragmentation categories also revealed that nearly5000 large patches (FC 1) together constituted 77 % of India’s forestedarea. The IPD for these large patches was also very low (median- 55m).This suggests that, in spite of ongoing degradation, with the propermanagement plan and mitigation strategies, connectivity between theselarge patches could be maintained and restored with minimal cost andresources. Identifying the critical points of animal crossing across thesepatches and providing effective structures across these linear intrusionswould greatly help in restoring connectivity across forests (W. F.Laurance et al., 2009; Lesbarrères and Fahrig, 2012).

There was a 70 % reduction in the number of patches which arelarger than 10,000 km2 in size compared to a scenario that excluded alllinear infrastructure. Furthermore, forest habitats of size greater than10 km2 were found to be more vulnerable to size reduction. We foundthat the mean patch size was only 2.2 km2 with more than 90 % of

patches being smaller than 1 km2 in size. Studies in South Ecuador alsoreported a smaller mean patch size, which decreased from 15.1 km2 in1976 to only 1.4 km2 (Tapia-armijos et al., 2015). Smaller patchessustain fewer species than larger or contiguous habitats and the ex-tinction threat to a species increases in smaller fragments (Estrada-Villegas et al., 2010; Gibson et al., 2011; Gibson et al., 2013; Melo et al.,2010). Recent studies have suggested that smaller PAs are dis-proportionately important for connectivity (Dutta et al., 2016) andsmall isolated PAs have elevated the risk of tiger extinction (Thatteet al., 2018). Large patches could help to increase the probability oflong-term persistence of large mammals by sustaining meta-populations(Wikramanayake et al., 2004). Studies also suggest that the size offorest habitats required for conserving genetic diversity could be 15folds larger than those needed to safeguard species numbers (Struebig,2011). Hence, protection of identified larger patches and restoringconnectivity should be a conservation priority.

The reduced habitat size and increased forest edges may lead to anescalation of human-wildlife conflict (Fredriksson and Fredrikkson,2005; Gurung et al., 2008; Michalski et al., 2006). The PA network inIndia covers only 5% of the total landscape (Rangarajan andShahabuddin, 2006), but has played an important role in conserving thehabitat of several species in India (Das et al., 2006; Wikramanayakeet al., 2004). However, 70 % of the assessed PAs were part of large andrelatively intact forest habitats, pointing to their role in decreasingforest degradation in India. Global studies also show that PAs havesignificantly lower rates of forest clearance and have played a majorrole in conserving tropical biodiversity (Bruner et al., 2001; Nagendra,2008). Declaring some of the identified larger patches as PAs could be

Fig. 5. Large, intact forest patches (FC 1) formed 75 % of total forested area in India.

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

8

an effective way of ceasing fragmentation and maintaining con-nectivity.

The Western Ghats and Central India are two important conserva-tion landscapes critical for the survival of several threatened mammalsincluding tiger Panthera tigris, Asian elephant Elephas maximus, leopardPanthera pardus, dhole Cuon alpinus, gaur Bos gaurus, sambar Rusaunicolor, lion-tailed macaque Macaca silenus etc. Results suggest thatoverall Central India has a greater proportion of forest area (42 %)comprised of larger patches (> 1000 km2), while only 29 % of forestarea in the Western Ghats was contributed by patches larger than1000 km2 in size. There were fewer major roads in Central Indiacompared to the Western Ghats, and this might be a reason for morenumber of larger patches in Central India compared to the Western

Ghats. However, inter-patch distance and perforation were higher forCentral India when compared to the Western Ghats, which is an in-dication of poor quality of the existing forests. We suggest that thepriority in Central India shall be to restore connectivity between largepatches and minimize forest diversion for non-forestry activities such asmining. The integrity of smaller intervening patches shall also be con-sidered while planning infrastructure as small patches have been foundto facilitate movement of dispersing individuals between larger patches(Thatte et al., 2018). Based on the fragmentation indices for the Wes-tern Ghats, we suggest that the priority in this landscape should be toavoid further fragmentation of patches and explore options of estab-lishing infrastructure along the existing features. In both of the land-scapes, the majority of the forested habitat was outside the PA network.

Fig. 6. (a & b) A comparison of distribution of forest cover across different patch sizes suggested that in the Central Indian landscape ∼42 % of total forest cover iscomprised of patches of size> 1000 km2, whereas only 29 % of forest cover in the Western Ghats landscape is comprised of patches of size> 1000 km2; (c & d).Median amount of perforation for larger patches of size> 250 km2 was higher in Central India than in the Western Ghats landscape.

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

9

Some of these forest patches outside PAs have been identified as crucialcorridors and critical links for biodiversity conservation (Das et al.,2006; Joshi et al., 2013; Sharma et al., 2013; Thatte et al., 2018).

The forests in the hill ranges of the Western Ghats run parallel to thewest coast of India. There are opportunities to establish linear infra-structure projects that connect north-south parts of this landscape, byaligning them along the coast or on the leeward side of the hill rangewithout fragmenting the forest patches. However, the establishment ofan east-west linear infrastructure project is feasible only along thenatural breaks in the hills or such projects will necessarily result insome amount forest fragmentation. Furthermore, the costs associatedwith establishing E–W connecting infrastructures through the hillyterrain in the Western Ghats are very high and hence, the options ofbundling several linear projects together along the natural breaks couldbe explored in this landscape. Unlike the Western Ghats, Central Indialandscape is in the heart of the country and crucial for establishingNorth-South and East-West linear infrastructure network. In the currentcontext, where there is limited forest cover, we recommend that in-frastructure projects should not be established through the existingforests and when inevitable, proper mitigation strategies to avoid ne-gative impacts on forest connectivity shall be incorporated in the pro-ject development plans. A more rational development planning wouldbe to connect larger numbers of villages or people while safeguardingforests rather than to establish the shortest routes that would destroyforests, biodiversity, and ecosystem services.

5. Conclusion

India is one of the fastest growing economies in the world. Its in-frastructure network is undergoing great expansion and up-gradation. Itis essential to integrate biodiversity conservation into this infra-structure development to achieve sustainable development in India. Wefound that States in the north-east had the highest forest cover and leastinfrastructure (Supplementary Table 3). Providing more infrastructurefacilities in these areas is an identified economic priority for the Indiangovernment. In this context, there is much that can be learned from acareful study of the impact of past infrastructure development in re-gions of India such as the Western Ghats and Central India. For ex-ample, there has been a substantial number of forest clearance appli-cations for linear projects submitted to the MoEFCC in these states(Supplementary Table 3, MoEFCC 2017). Hence, conservation priorityshould be towards identifying forest habitats that would be affected bylinear intrusions and developing effective mitigation strategies to pre-serve connectivity across the habitats. Several large-scale projects likeriver-interlinking (National Water Development Agency 2014),Bharatmala Pariyojana (National portal of India 2018), along withother linear projects could potentially increase fragmentation and affectconnectivity across landscapes. The increasing trend of fragmentationwould severely affect biodiversity by encouraging the spread of in-vasive species along the edges (Bustamante et al., 2003). Linear infra-structure can act as barriers to animal movement which in turn hasnegative demographic and genetic consequences that can result in localextinctions (Shepard et al., 2008). Fragmentation due to infrastructuredevelopments could be minimized if these structures could be rerouted,bundled, or effective mitigation strategies are adopted. Data drivenmodeling approaches could be used to quantify the potential effects ofthese structures on fragmentation and connectivity.

Our study and its results provide information to guide ‘where’ and'how' future infrastructure development activities could be undertakenwith the optimal balancing of development and biodiversity conserva-tion. Forest patches in fragmentation category 1 are large and relativelyintact, and hence, future development plans should avoid routingstructures through these patches. We are sharing this spatial datathrough a web-portal for the public to use (https://indiaunderconstruction.com/). Access to this data will enable localstakeholders in infrastructure projects to make informed decisions

leading to develop optimal local management plans for them. Althoughwe address only the issue of fragmentation within forests in this study,the techniques we have developed and presented here can be usefullyapplied to mitigate fragmentation problems in other fragile ecosystemssuch as tropical grasslands and savannas. Infrastructure expansioncomparable to that in India is also rapidly occurring in many tropicalcountries (Venter et al., 2016) driven by the societal need for economicdevelopment. However, we believe knowledge of current and futureimpacts of such infrastructure development on biodiversity can assistdeveloping societies to simultaneously align such development withconservation objectives, thus potentially presenting a win-win scenario.

CRediT authorship contribution statement

Rajat Nayak:Methodology, Data curation, Formal analysis, Writing- original draft. Krithi K. Karanth: Conceptualization, Funding acqui-sition, Supervision, Writing - review & editing. Trishna Dutta: Datacuration, Formal analysis, Writing - review & editing. Ruth Defries:Conceptualization, Funding acquisition, Writing - review & editing. K.Ullas Karanth: Conceptualization, Funding acquisition, Writing - re-view & editing. Srinivas Vaidyanathan: Conceptualization, Fundingacquisition, Supervision, Methodology, Data curation, Formal analysis,Writing - review & editing.

Acknowledgement

This work was carried out with funding from the Science for Natureand People Program (SNAPP) for the project on “Connectivity acrossthe landscape: Strategies to meet needs for infrastructure and wildlife inIndia.” We would like to thank Dr. Jagdish Krishnaswamy for his va-luable comments. We thank an anonymous reviewer for thoughtfulcomments on an earlier version of this manuscript.

Appendix A. Supplementary data

Supplementary material related to this article can be found, in theonline version, at doi:https://doi.org/10.1016/j.landusepol.2020.104619.

References

Achard, F., Stibig, H.J., Eva, H., Mayaux, P., 2002. Tropical forest cover monitoring in thehumid tropics – TREES project. Trop. Ecol. 43 (1), 9–20.

Baskaran, N., 2013. An overview of Asian Elephants in the Western Ghats, southern India:implications for the conservation of Western Ghats ecology. J. Threatened Taxa 5(October), 4854–4870.

Baskaran, N., Balasubramanian, M., Swaminathan, S., Desai, A.A., 1995. Home Range ofElephants in the Nilgiri Biosphere Reserve, South India. A Week With Elephants.Bombay Natural History Society and Oxford University Press, Bombay, pp. 296–313.

Benítez-López, A., Alkemade, R., Verweij, P.A., 2010. The impacts of roads and otherinfrastructure on mammal and bird populations: a meta-analysis. Biol. Conserv. 143(6), 1307–1316. https://doi.org/10.1016/j.biocon.2010.02.009.

Bhattacharya, M., Primack, R.B., Gerwein, J., 2002. Are roads and railroads barriers tobumblebee movement in a temperate suburban conservation area? Biol. Conserv. 109(1), 37–45. https://doi.org/10.1016/S0006-3207(02)00130-1.

Bruner, A.G., Gullison, R.E., Rice, R.E., Fonseca, G.A.B., 2001. Effectiveness of parks inprotecting tropical biodiversity. Science 291 (January), 125–128.

Bustamante, R.O., Serey, I.A., Pickett, S.T.A., 2003. Forest Fragmentation, PlantRegeneration and Invasion Processes Across Edges in Central Chile. In HowLandscapes Change. Springer, pp. 145–160.

Coffin, A.W., 2007. From roadkill to road ecology: a review of the ecological effects ofroads. J. Transp. Geogr. 15 (5), 396–406. https://doi.org/10.1016/j.jtrangeo.2006.11.006.

Das, A., Krishnaswamy, J., Bawa, K.S., Kiran, M.C., Srinivas, V., Kumar, N.S., Karanth,K.U., 2006. Prioritisation of conservation areas in the Western Ghats, India. Biol.Conserv. 133 (1), 16–31. https://doi.org/10.1016/j.biocon.2006.05.023.

Develey, P.F., Stouffer, P.C., 2001. Effects of roads on movements by understory birds inmixed-species flocks in Central Amazonian Brazil. Conserv. Biol. 15 (5), 1416–1422.https://doi.org/10.1046/j.1523-1739.2001.00170.x.

Dutta, T., Sharma, S., McRae, B.H., Roy, P.S., DeFries, R., 2016. Connecting the dots:mapping habitat connectivity for tigers in central India. Reg. Environ. Change 16 (1),53–67.

Dutta, T., Sharma, S., DeFries, R., 2018. Targeting restoration sites to improve

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

10

https://indiaunderconstruction.com/https://indiaunderconstruction.com/https://doi.org/10.1016/j.landusepol.2020.104619https://doi.org/10.1016/j.landusepol.2020.104619http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0005http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0005http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0010http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0010http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0010http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0015http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0015http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0015https://doi.org/10.1016/j.biocon.2010.02.009https://doi.org/10.1016/S0006-3207(02)00130-1http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0030http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0030http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0035http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0035http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0035https://doi.org/10.1016/j.jtrangeo.2006.11.006https://doi.org/10.1016/j.jtrangeo.2006.11.006https://doi.org/10.1016/j.biocon.2006.05.023https://doi.org/10.1046/j.1523-1739.2001.00170.xhttp://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0055http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0055http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0055http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0060

connectivity in a tiger conservation landscape in India. PeerJ 6, e5587.Estrada-Villegas, S., Meyer, C.F.J., Kalko, E.K.V., 2010. Effects of tropical forest frag-

mentation on aerial insectivorous bats in a land-bridge island system. Biol. Conserv.143 (3), 597–608. https://doi.org/10.1016/j.biocon.2009.11.009.

Fahrig, L., 2002. Effect of habitat fragmentation on the extinction threshold: a synthesis.Ecol. Appl. 12 (2), 346–353. https://doi.org/10.1890/1051-0761.

Forman, R.T.T., 2003. Road Ecology: Science and Solutions. Island Press.Forman, R.T.T., Alexander, L.E., 1998. Roads and their major ecological effects. Annu.

Rev. Ecol. Syst. 29 (1), 207–231. https://doi.org/10.1146/annurev.ecolsys.29.1.207.Forman, R.T., Friedman, D.S., Fitzhenry, D., Martin, J.D., Chen, A.S., Alexander, L.E.,

1997. Ecological effects of roads: toward three summary indices and an overview forNorth America. Habitat Fragment. Infrastruct. – Proc. 21, 40–54.

Fredriksson, G., Fredrikkson, G., 2005. Human-sun bear conflicts in East Kalimantan,Indonesian Borneo. Ursus 16 (1), 130–137. https://doi.org/10.2192/1537-6176(2005)016[0130:hbciek]2.0.co;2.

Geist, H.J., Lambin, E.F., 2002. Proximate causes and underlying driving forces of tropicaldeforestation. BioScience 52 (2), 143. https://doi.org/10.1641/0006-3568(2002)052[0143:PCAUDF]2.0.CO;2.

Gelbard, J.L., Harrison, S., 2003. Roadless habitats as refuges for native grasslands: in-teractions with soil, aspect, and grazing. Ecol. Appl. 13 (2), 404–415. https://doi.org/10.1890/1051-0761.

Geneletti, D., 2004. Using spatial indicators and value functions to assess ecosystemfragmentation caused by linear infrastructures. Int. J. Appl. Earth Obs. Geoinf. 5 (1),1–15. https://doi.org/10.1016/j.jag.2003.08.004.

Gibson, L., Lee, T.M., Koh, L.P., Brook, B.W., Gardner, T.A., Barlow, J., et al., 2011.Primary forests are irreplaceable for sustaining tropical biodiversity. Nature 478(7369), 378–381. https://doi.org/10.1038/nature10425.

Gibson, L., Lynam, A.J., Bradshaw, C.J.A., He, F., Bickford, D.P., 2013. Near-completeextinction of native small mammal fauna 25 years after forest fragmentation. Science45 (September), 1508–1510.

Girardet, X., Foltête, J., Clauzel, C., 2013. Designing a graph-based approach to landscapeecological assessment of linear infrastructures. Environ. Impact Assess. Rev. 42,10–17. https://doi.org/10.1016/j.eiar.2013.03.004.

Goosem, M., 1997. Internal Fragmentation: the Effects of Roads, Highways, andPowerline Clearings on Movements and Mortality of Rainforest Vertebrates. TropicalForest Remnants: Ecology, Management, and Conservation of FragmentedCommunities. University of Chicago Press, Chicago, pp. 241–255.

Goosem, M., 2007. Fragmentation impacts caused by roads through rainforests. Curr. Sci.93 (11), 1587–1595. Retrieved from. http://www.jstor.org/stable/24099089.

Gurung, B., Smith, J.L.D., McDougal, C., Karki, J.B., Barlow, A., 2008. Factors associatedwith human-killing tigers in Chitwan National Park, Nepal. Biol. Conserv. 141 (12),3069–3078. https://doi.org/10.1016/j.biocon.2008.09.013.

Hawbaker, T.J., Adeloff, V.C.R., Layton, M.K.C., Ammer, R.B.H., 2006. Road develop-ment, housing growth, and landscape fragmentation in northern Wisconsin:1937–1999. Ecol. Appl. 16 (3), 1222–1237.

Hill, J.K., Gray, M.A., Khen, C.V., Benedick, S., Tawatao, N., Hamer, K.C., 2011.Ecological impacts of tropical forest fragmentation: How consistent are patterns inspecies richness and nestedness? Philos. Trans. Biol. Sci. 366 (1582), 3265–3276.

Jaarsma, C.F., 2006. Flattened fauna and mitigation: traffic victims related to road,traffic, vehicle, and species characteristics. Transp. Res. D Transp. Environ. 11,264–276. https://doi.org/10.1016/j.trd.2006.05.001.

Joshi, A., Vaidyanathan, S., Mondol, S., Edgaonkar, A., Ramakrishnan, U., 2013.Connectivity of tiger (Panthera tigris) populations in the human-influenced forestmosaic of Central India. PLoS One 8 (11). https://doi.org/10.1371/journal.pone.0077980.

Karlson, M., Mörtberg, U., 2015. A spatial ecological assessment of fragmentation anddisturbance effects of the Swedish road network. Landsc. Urban Plan. 134, 53–65.https://doi.org/10.1016/j.landurbplan.2014.10.009.

Kerley, L.L., Goodrich, J.M., Miquelle, D.G., Smirnov, E.N., Quigley, H.B., Hornocker,M.G., 2002. Effects of roads and human disturbance on Amur tigers. Conserv. Biol. 16(1), 97–108. https://doi.org/10.1046/j.1523-1739.2002.99290.x.

Kociolek, A.V., Clevenger, A.P., St. Clair, C.C., Proppe, D.S., 2011. Effects of road net-works on bird populations. Conserv. Biol. 25 (2). https://doi.org/10.1111/j.1523-1739.2010.01635.x. no-no.

Laurance, W.F., 2015. Emerging threats to tropical forests. Ann. Mo. Bot. Gard. 100 (3),159–169. https://doi.org/10.3417/2011087.

Laurance, W.F., Bierregaard, R.O., 1997. Tropical Forest Remnants: Ecology,Management, and Conservation of Fragmented Communities. University of ChicagoPress.

Laurance, S.G.W., Gomez, M.S., 2005. Clearing width and movements of understoryrainforest birds. Biotropica 37 (1), 149–152. https://doi.org/10.1111/j.1744-7429.2005.04099.x.

Laurance, W.F., Goosem, M., Laurance, S.G.W., 2009. Impacts of roads and linear clear-ings on tropical forests. Trends Ecol. Evol. 24 (12), 659–669. https://doi.org/10.1016/j.tree.2009.06.009.

Laurance, W.F., Clements, G.R., Sloan, S., O’Connell, C.S., Mueller, N.D., Goosem, M.,et al., 2014. A global strategy for road building. Nature 513, 229. https://doi.org/10.1038/nature13717. Retrieved from.

Laurance, W.F., Camargo, J.L.C., Fearnside, P.M., Lovejoy, T.E., Williamson, G.B.,Mesquita, R.C.G., et al., 2017. An Amazonian rainforest and its fragments as a la-boratory of global change. Biol. Rev. https://doi.org/10.1111/brv.12343.

Lesbarrères, D., Fahrig, L., 2012. Measures to reduce population fragmentation by roads:what has worked and how do we know? Trends Ecol. Evol. 27 (7), 374–380. https://doi.org/10.1016/j.tree.2012.01.015.

Lovejoy, T.E., Bierregaard, R.O.J., Rylands, A.B., Malcolm, J.R., Quintela, C.E., Harper,L.H., et al., 1986. Edge and other effects of isolation on Amazon forest fragments.

Conserv. Biol.: Sci. Scarcity Diversity.Mancebo Quintana, S., Martín Ramos, B., Casermeiro Martínez, M.A., Otero Pastor, I.,

2010. A model for assessing habitat fragmentation caused by new infrastructures inextensive territories – evaluation of the impact of the Spanish strategic infrastructureand transport plan. J. Environ. Manage. 91 (5), 1087–1096. https://doi.org/10.1016/j.jenvman.2009.12.013.

Mayaux, P., Holmgren, P., Achard, F., Eva, H., Stibig, H.-J., Branthomme, A., 2005.Tropical forest cover change in the 1990s and options for future monitoring. Philos.Trans. Biol. Sci. 360 (1454), 373–384.

Melo, F.P.L., Mart, E., Ben, J., Ceballos, G., 2010. Forest fragmentation reduces recruit-ment of large-seeded tree species in a semi-deciduous tropical forest of southernMexico. J. Trop. Ecol. 35–43. https://doi.org/10.1017/S0266467409990435.

Michalski, F., Boulhosa, R.L.P., Faria, A., Peres, C.A., 2006. Human-wildlife conflicts in afragmented Amazonian forest landscape: determinants of large felid depredation onlivestock. Anim. Conserv. 9 (2), 179–188. https://doi.org/10.1111/j.1469-1795.2006.00025.x.

Miles, L., Newton, A.C., DeFries, R.S., Ravilious, C., May, I., Blyth, S., et al., 2006. Aglobal overview of the conservation status of tropical dry forests. J. Biogeogr. 33 (3),491–505. https://doi.org/10.1111/j.1365-2699.2005.01424.x.

Miller, J.R., Joyce, L.A., Knight, R.L., King, R.M., 1996. Forest roads and landscapestructure in the southern Rocky Mountains. Landsc. Ecol. 11 (2), 115–127.

Mittermeier, R.A., Robles, G.P., Michael, H., John, P., Thomas, B., Mittermeier, C.G.,Lamoreux, J.L., Fonseca, G., 2004. Hotspots Revisited: Earth’s Biologically Richestand Most Endangered Terrestrial Ecoregions. CEMEX, Mexico.

Müllner, D., 2013. Fastcluster: fast hierarchical, agglomerative clustering routines for Rand python. J. Stat. Softw. 53 (9), 1–18. URL. http://www.jstatsoft.org/v53/i09/.

Nagendra, H., 2008. Do parks work? Impact of Protected Areas on land cover clearing.Ambio A J. Hum. Environ. 37 (5), 330–337. https://doi.org/10.1579/06-R-184.1.

Opdam, P., Foppen, R., Vos, C., 2001. Bridging the gap between ecology and spatialplanning in landscape ecology. Landsc. Ecol. 16 (8), 767–779.

Pekel, J.F., Cottam, A., Gorelick, N., Belward, A.S., 2016. High-resolution mapping ofglobal surface water and its long-term changes. Nature 540 (7633), 418–422. https://doi.org/10.1038/nature20584.

R Core Team, 2018. R: a Language and Environment for Statistical Computing. URL. RFoundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/.

Ramachandran, R.M., Roy, P.S., Chakravarthi, V., Sanjay, J., Joshi, P.K., 2018. Long-termland use and land cover changes (1920–2015) in Eastern Ghats, India: pattern ofdynamics and challenges in plant species conservation. Ecol. Indic. 85 (March 2017),21–36. https://doi.org/10.1016/j.ecolind.2017.10.012.

Rangarajan, M., Shahabuddin, G., 2006. Displacement and relocation from protectedareas: towards a biological and historical synthesis. Conserv. Soc. 4 (3), 359–378.

Reddy, C.S., Sreelekshmi, S., Jha, C.S., Dadhwal, V.K., 2013. National assessment of forestfragmentation in India: landscape indices as measures of the effects of fragmentationand forest cover change. Ecol. Eng. 60, 453–464. https://doi.org/10.1016/j.ecoleng.2013.09.064.

Reed, R.A., Johnson-Barnard, J., Baker, W.L., 1996. Contribution of roads to forestfragmentation in the Rocky Mountains. Conserv. Biol. 10 (4), 1098–1106. https://doi.org/10.1046/j.1523-1739.1996.10041098.x.

Riley, S.P.D., Pollinger, J.P., Sauvajot, R.M., York, E.C., Bromley, C., Fuller, T.K., Wayne,R.K., 2006. A southern California freeway is a physical and social barrier to gene flowin carnivores. Mol. Ecol. 15 (7), 1733–1741. https://doi.org/10.1111/j.1365-294X.2006.02907.x.

Root‐Bernstein, M., Svenning, J.C., 2018. Human paths have positive impacts on plantrichness and diversity: A meta‐analysis. Ecol. Evol. 8 (22), 11111–11121.

Sanderson, E.W., Forrest, J., Loucks, C., Ginsberg, J., Dinerstein, E., Seidensticker, J.,et al., 2006. Setting priorities for tiger conservation: 2005-2015. In: Tilson, R., Nyhus,P.J. (Eds.), Tigers of the World. Elsevier, New York – Washington, D.C.

Schowalter, T.D., 2016. Insect Ecology: an Ecosystem Approach. Academic Press.Sehgal, J., Mandal, D.K., Mandal, C., Vadivelu, S., 1992. Agro-ecological Regions of India.

National Bureau of Soil Survey and Land Use Planning, Nagpur, India.Seiler, A., Eriksson, I., 1997. New approaches for ecological consideration in Swedish

road planning New approaches for ecological consideration in Swedish road plan-ning. In: Canters, K., Piepers, A., Hendriks-Heersma, A. (Eds.), HabitatFragmentation, Infrastructure and the Role of Ecological Engineering. Delft, TheNetherlands: Maastricht & DenHague, pp. 253–264 1995.

Serrano, M., Sanz, L., Puig, J., Pons, J., 2002. Landscape fragmentation caused by thetransport network in Navarra (Spain) Two-scale analysis and landscape integrationassessment. Landsc. Urban Plan. 58 (2–4), 113–123. https://doi.org/10.1016/S0169-2046(01)00214-6.

Sharma, S., Dutta, T., Wood, C., Panwar, H.S., Seidensticker, J., 2013. Forest corridorsmaintain historical gene flow in a tiger metapopulation in the highlands of centralIndia. Proc. R. Soc. B: Biol. Sci. 280 (1767), 20131506.

Shepard, D.B., Kuhns, A.R., Dreslik, M.J., Phillips, C.A., 2008. Roads as barriers to animalmovement in fragmented landscapes. Anim. Conserv. 11 (4), 288–296. https://doi.org/10.1111/j.1469-1795.2008.00183.x.

Strand, O., 2004. Effects of infrastructure on migration. J. Wildl. Manage. 68 (1),101–108. https://doi.org/10.2193/0022-541X.

Struebig, M.J., 2011. Parallel declines in species and genetic diversity in tropical forestfragments. Ecol. Lett. 14 (6), 582–590. https://doi.org/10.1111/j.1461-0248.2011.01623.x.

Tapia-armijos, M.F., Homeier, J., Espinosa, C.I., 2015. Deforestation and forest frag-mentation in South Ecuador Since the 1970s – Losing a hotspot of biodiversity. PLoSOne 10 (9), 1–18. https://doi.org/10.5061/dryad.32451.

Thatte, P., Joshi, A., Vaidyanathan, S., Landguth, E., Ramakrishnan, U., 2018.Maintaining tiger connectivity and minimizing extinction into the next century: in-sights from landscape genetics and spatially-explicit simulations. Biol. Conserv. 218

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

11

http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0060https://doi.org/10.1016/j.biocon.2009.11.009https://doi.org/10.1890/1051-0761http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0075https://doi.org/10.1146/annurev.ecolsys.29.1.207http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0085http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0085http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0085https://doi.org/10.2192/1537-6176(2005)016[0130:hbciek]2.0.co;2https://doi.org/10.2192/1537-6176(2005)016[0130:hbciek]2.0.co;2https://doi.org/10.1641/0006-3568(2002)052[0143:PCAUDF]2.0.CO;2https://doi.org/10.1641/0006-3568(2002)052[0143:PCAUDF]2.0.CO;2https://doi.org/10.1890/1051-0761https://doi.org/10.1890/1051-0761https://doi.org/10.1016/j.jag.2003.08.004https://doi.org/10.1038/nature10425http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0115http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0115http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0115https://doi.org/10.1016/j.eiar.2013.03.004http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0125http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0125http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0125http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0125http://www.jstor.org/stable/24099089https://doi.org/10.1016/j.biocon.2008.09.013http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0140http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0140http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0140http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0145http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0145http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0145https://doi.org/10.1016/j.trd.2006.05.001https://doi.org/10.1371/journal.pone.0077980https://doi.org/10.1371/journal.pone.0077980https://doi.org/10.1016/j.landurbplan.2014.10.009https://doi.org/10.1046/j.1523-1739.2002.99290.xhttps://doi.org/10.1111/j.1523-1739.2010.01635.xhttps://doi.org/10.1111/j.1523-1739.2010.01635.xhttps://doi.org/10.3417/2011087http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0180http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0180http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0180https://doi.org/10.1111/j.1744-7429.2005.04099.xhttps://doi.org/10.1111/j.1744-7429.2005.04099.xhttps://doi.org/10.1016/j.tree.2009.06.009https://doi.org/10.1016/j.tree.2009.06.009https://doi.org/10.1038/nature13717https://doi.org/10.1038/nature13717https://doi.org/10.1111/brv.12343https://doi.org/10.1016/j.tree.2012.01.015https://doi.org/10.1016/j.tree.2012.01.015http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0210http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0210http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0210https://doi.org/10.1016/j.jenvman.2009.12.013https://doi.org/10.1016/j.jenvman.2009.12.013http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0220http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0220http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0220https://doi.org/10.1017/S0266467409990435https://doi.org/10.1111/j.1469-1795.2006.00025.xhttps://doi.org/10.1111/j.1469-1795.2006.00025.xhttps://doi.org/10.1111/j.1365-2699.2005.01424.xhttp://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0240http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0240http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0245http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0245http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0245http://www.jstatsoft.org/v53/i09/https://doi.org/10.1579/06-R-184.1http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0260http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0260https://doi.org/10.1038/nature20584https://doi.org/10.1038/nature20584https://www.R-project.org/https://doi.org/10.1016/j.ecolind.2017.10.012http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0280http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0280https://doi.org/10.1016/j.ecoleng.2013.09.064https://doi.org/10.1016/j.ecoleng.2013.09.064https://doi.org/10.1046/j.1523-1739.1996.10041098.xhttps://doi.org/10.1046/j.1523-1739.1996.10041098.xhttps://doi.org/10.1111/j.1365-294X.2006.02907.xhttps://doi.org/10.1111/j.1365-294X.2006.02907.xhttp://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0300http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0300http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0305http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0305http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0305http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0310http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0315http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0315http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0320http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0320http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0320http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0320http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0320https://doi.org/10.1016/S0169-2046(01)00214-6https://doi.org/10.1016/S0169-2046(01)00214-6http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0330http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0330http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0330https://doi.org/10.1111/j.1469-1795.2008.00183.xhttps://doi.org/10.1111/j.1469-1795.2008.00183.xhttps://doi.org/10.2193/0022-541Xhttps://doi.org/10.1111/j.1461-0248.2011.01623.xhttps://doi.org/10.1111/j.1461-0248.2011.01623.xhttps://doi.org/10.5061/dryad.32451

(December 2017), 181–191. https://doi.org/10.1016/j.biocon.2017.12.022.Uddin, M., 2017. Assessing threats to birds from power-lines in Thar with special em-

phasis on Great Indian Bustard. University of Kota, Rajasthan.Uezu, A., Metzger, J.P., Vielliard, J.M., 2005. Effects of structural and functional con-

nectivity and patch size on the abundance of seven Atlantic Forest bird species. Biol.Conserv. 123 (4), 507–519.

Venter, O., Sanderson, E.W., Magrach, A., Allan, J.R., Beher, J., Jones, K.R., et al., 2016.Sixteen years of change in the global terrestrial human footprint and implications forbiodiversity conservation. Nat. Commun. 7, 1–11. https://doi.org/10.1038/ncomms12558.

Ward, J.H., 1963. Hierarchical grouping to optimize an objective function. J. Am. Stat.Assoc. 58 (301), 236–244.

Webb, S., 2013. Corridors to Extinction and the Australian Megafauna. Newnes.WII, 2016. Eco-friendly Measures to Mitigate Impacts of Linear Infrastructure on Wildlife.

Dehradun, India. Retrieved from. http://www.moef.gov.in/sites/default/files/Invitingcommnets%26suggestions.pdf.

Wikramanayake, E., Knight, M.M.C., Dinerstein, E., Joshi, A., Gurung, B., Smith, D., 2004.Designing a conservation landscape for tigers in human-dominated environments.Conserv. Biol. 18 (3), 839–844.

R. Nayak, et al. Land Use Policy xxx (xxxx) xxxx

12

https://doi.org/10.1016/j.biocon.2017.12.022http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0360http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0360http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0365http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0365http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0365https://doi.org/10.1038/ncomms12558https://doi.org/10.1038/ncomms12558http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0375http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0375http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0380http://www.moef.gov.in/sites/default/files/Invitingcommnets%26suggestions.pdfhttp://www.moef.gov.in/sites/default/files/Invitingcommnets%26suggestions.pdfhttp://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0390http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0390http://refhub.elsevier.com/S0264-8377(18)31322-X/sbref0390

Bits and pieces: Forest fragmentation by linear intrusions in IndiaIntroductionMethodsDatasetsForest cover and infrastructure layersAdministrative and landscape boundaries

Preprocessing datasetsAnalysis

ResultsInfrastructure and patch indices at the national levelFragmentation category (FC)Status of existing Protected Area (PA) NetworkA comparison between the Western Ghats and Central India

DiscussionConclusionCRediT authorship contribution statementAcknowledgementSupplementary dataReferences